Research map / Large language models / Language, safety and interpretability
Unsupervised Bilingual Lexicon Induction from Mono-Lingual Multimodal Data
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| combine linguistic and localized visual features (ablation) | Bilingual lexicon induction model · BERGSMA German-English | MRR: 0.755 -> 0.819 | Helped |
| combine linguistic and localized visual features (ablation) | Bilingual lexicon induction model · MMID German-English | MRR: 0.467 -> 0.529 | Helped |
| combine linguistic and localized visual features (ablation) | Bilingual lexicon induction model · MMID French-English | MRR: 0.376 -> 0.494 | Helped |
| combine linguistic and localized visual features (ablation) | Bilingual lexicon induction model · MMID Japanese-English | MRR: 0.290 -> 0.469 | Helped |
| use localized visual features instead of global image features | CNN-mean vision-based baseline · BERGSMA German-English | MRR: 0.650 -> 0.762 | Helped |
| use the attention caption model instead of mean pooling (ablation) | Multilingual caption model · MMID German-English | MRR: 0.490 -> 0.529 | Helped |
| train a multilingual rather than separate monolingual caption model | Image captioning model · Multi30k French | CIDEr: 47.45 -> 50.67 | Helped |
| train a multilingual rather than separate monolingual caption model | Image captioning model · Multi30k English | BLEU4: 23.59 -> 23.39 | No clear effect |
| train a multilingual rather than separate monolingual caption model | Image captioning model · COCO and STAIR Japanese | BLEU4: 39.85 -> 39.81 | No clear effect |
Closest papers by meaning
- Semantic and Expressive Variations in Image Captions Across Languages (CVPR 2025 · similarity 0.90)
- Multilingual Diversity Improves Vision-Language Representations (NeurIPS 2024 · similarity 0.89)
- Embracing Language Inclusivity and Diversity in CLIP through Continual Language Learning (AAAI 2024 · similarity 0.87)
- Unified Lexical Representation for Interpretable Visual-Language Alignment (NeurIPS 2024 · similarity 0.87)
- Unified Language-Vision Pretraining in LLM with Dynamic Discrete Visual Tokenization (ICLR 2024 · similarity 0.87)
- CL2CM: Improving Cross-Lingual Cross-Modal Retrieval via Cross-Lingual Knowledge Transfer (AAAI 2024 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.